INDEX

Explanations

Code and data structures

np_max-act · gemini-2.0-flash

This neuron detects SQL/JSON string‐building patterns, especially the concatenation operator (||) and surrounding JSON delimiters.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

(marker

-0.06

_ALL

-0.06

↵

-0.06

ful

-0.06

_TILE

-0.06

атели

-0.06

-placeholder

-0.06

ishlist

-0.06

HEME

-0.06

ulatory

-0.06

POSITIVE LOGITS

 граж

0.06

orr

0.06

lias

0.06

�p

0.06

óz

0.06

 Riyadh

0.06

 dostal

0.06

 retro

0.06

 entitlement

0.06

شب

0.06

Activations Density 0.181%

Code and data structures

This neuron detects SQL/JSON string‐building patterns, especially the concatenation operator (||) and surrounding JSON delimiters.

No Comments

No Known Activations